Potentials and pitfalls in the analysis of bipartite networks to understand plant-microbe interactions in changing environments

Complex networks of interactions involving multiple plant and microbial species have consequences for ecosystem productivity and energy movement through higher trophic levels, but the sensitivity of these networks to environmental change is uncertain. The fast developing field focused on the construction of ecological networks involving multiple plant and microbial species allows the structure, complexity and robustness of these networks to be examined. However, the choice of methods used to generate data and construct networks can significantly influence network-level metrics. Here, we provide background and guidance on approaches for building and analysing bipartite networks of plant-microbe interactions. Researchers need to explore these methods (and their advantages and pitfalls) to create high quality and reproducible data that allow comparisons within and between systems. In particular, we highlight the significant advances of DNA-based approaches for characterising interactions between plants and microbes, while suggesting future work needed to help overcome their current limitations. As more studies embrace these approaches, these guidelines will allow robust characterisation and analysis of networks of plant and microbial interactions, and further enable our understanding of the influence of environmental change on species interactions and, ultimately, communities and ecosystem functioning.